Computers & Graphics-Uk最新文献

Representing historical figures on visual media has always been a crucial aspect of political communication in the ancient world, as it is in modern society. A great example comes from ancient Rome, when the emperor’s portraits were serially replicated on visual media to disseminate his image across the countries ruled by the Romans and to assert the power and authority that he embodied by making him universally recognizable. In particular, one of the most common media through which ancient Romans spread the imperial image was coinage, which showed a bi-dimensional projection of his portrait on the very low relief produced by the impression of the coin-die. In this work, we propose a new method that uses a multi-modal 2D and 3D approach to reconstruct the full portrait in the round of Roman emperors from their images adopted on ancient coins. A well-defined pipeline is introduced from the digitization of coins using 3D scanning techniques to the estimation of the 3D model of the portrait represented by a polygonal mesh. A morphable model trained on real 3D faces is exploited to infer the morphological (i.e., geometric) characteristics of the Roman emperor from the contours extracted from a coin portrait using a model fitting procedure. We present examples of face reconstruction of different emperors from coins produced in Rome as well as in the imperial provinces, which sometimes showed local variations of the official portraits centrally designed.

Inpainting techniques, rooted in ancient art restoration practices, have become essential tools for digital image editing in modern contexts. Despite their widespread applications across diverse domains, the rapid advance of inpainting methodologies has highlighted the need for comprehensive reviews to document progress and identify areas for deeper investigation. Although there are many works in literature describing the state of the art regarding inpainting methods, algorithms, and technologies, many of them are presented lacking methodological rigor, which compromises the reliability and validity of their conclusions. In light of the wide literature about inpainting, this tertiary review aims to systematically identify their main techniques, recurring challenges, and applications through the perspective of secondary studies, providing a helpful background for new researchers. Our findings are based on an analysis of 45 reviews, where one of the major issues observed was the lack of standardization in the classification of methods, and to address this, we provide a concise and clear classification. Furthermore, we present a summary of the most commonly used metrics and a discussion of the main shortcomings and applications, which extend beyond digital image restoration to include medical imaging, three-dimensional restoration, cultural heritage preservation, and more. While inpainting poses challenges, this review aims to inspire further exploration and advancement in the field by providing a comprehensive overview of inpainting research.

Positive airway pressure (PAP) therapy refers to sleep disordered breathing treatment that uses a stream of compressed air to support the airway during sleep. Even though the use of PAP therapy has been shown to be effective in improving the symptoms and quality of life, many patients are intolerant of the treatment due to poor mask fit. In this paper, our goal is to develop a computational approach for designing custom-fit PAP masks such that they can achieve better mask fit performance in terms of mask leakage and comfort. Our key observation is that a custom-fit PAP mask should fit a patient’s face in its deformed state instead of in its rest state since the PAP mask cushion undergoes notable deformation before reaching an equilibrium state during PAP therapy. To this end, we compute the equilibrium state of a mask cushion using the finite element method, and quantitatively measure the leakage and comfort of the mask cushion in this state. We further optimize the mask cushion geometry to minimize the two measures while ensuring that the cushion can be easily fabricated with molding. We demonstrate the effectiveness of our computational approach on a variety of face models and different types of PAP masks. Experimental results on real subjects show that our designed custom-fit PAP masks are able to achieve better mask fit performance than a generic PAP mask and custom-fit PAP masks designed by a state-of-the-art approach.

Given an input query, a search and retrieval system fetches relevant information from a dataset. In the Engineering domain, such a system is beneficial for tasks such as design reuse. A two-dimensional (2D) sketch is more conducive for an end user to give as a query than a three-dimensional (3D) object. Such query sketches, nevertheless, will inevitably contain defects like incomplete lines, mesh lines, overdrawn areas, missing areas, etc. Since a retrieval system’s results are only as good as the query, it is necessary to improve the query sketches.

In this paper, the problem of transforming a defective CAD sketch into a defect-free sketch is addressed using Generative Adversarial Networks (GANs), which, to the best of our knowledge, has not been investigated before. We first create a dataset of 534 hand-drawn sketches by tracing the boundaries of images of CAD models. We then pair the corrected sketches with their corresponding defective sketches and use them for training a C-WGAN (Conditional Wasserstein Generative Adversarial Network), called SketchCleanGAN. We model the transformation from defective to defect-free sketch as a factorization of the defective input sketch and then translate it to the space of defect-free sketch. We propose a three-branch strategy to this problem. Ablation studies and comparisons with other state-of-the-art techniques demonstrate the efficacy of the proposed technique. Additionally, we also contribute to a dataset of around 58000 improved sketches using the proposed framework.

Classic generalized subdivision, such as Catmull–Clark subdivision, as well as recent subdivision algorithms for high-quality surfaces, rely on slower convergence towards extraordinary points for mesh nodes surrounded by $n>4$ quadrilaterals. Slow convergence corresponds to a contraction-ratio of $\lambda >0.5$. To improve shape, prevent parameterization discordant with surface growth, or to improve convergence in isogeometric analysis near extraordinary points, a number of algorithms explicitly adjust $\lambda$ by altering refinement rules. However, such tuning of $\lambda$ has so far led to poorer surface quality, visible as uneven distribution or oscillation of highlight lines. The recent Quadratic-Attraction Subdivision (QAS) generates high-quality, bounded curvature surfaces based on a careful choice of quadratic expansion at the central point and, just like Catmull–Clark subdivision, creates the control points of the next subdivision ring by matrix multiplication. But QAS shares the contraction-ratio ${\lambda }_{CC}>1/2$ of Catmull–Clark subdivision when $n>4$. For $n=5,\dots ,10$, QAS${}_{+}$ improves the convergence to the uniform $\lambda =\frac{1}{2}$ of binary domain refinement and without sacrificing surface quality compared to QAS.

We propose a deep learning approach for inpainting holes in digital models of fabric surfaces. Leveraging the developable nature of fabric surfaces, we flatten the area surrounding the holes with minor distortion and regularly sample it to obtain a discrete 2D map of the 3D embedding, with an indicator mask outlining holes locations. This enables the use of a standard 2D convolutional neural network to inpaint holes given the 3D positioning of the surface. The provided neural architecture includes an attention mechanism to capture long-range relationships on the surface. Finally, we provide ScarfFolds, a database of folded fabrics patches with varying complexity, which is used to train our convolutional network in a supervised manner. We successfully tested our approach on various examples and illustrated that previous 3D deep learning approaches suffer from several issues when applied to fabrics. Also, our method allows the users to interact with the construction of the inpainted surface. The editing is interactive and supports many tools like vertex grabbing, drape twisting or pinching.

Three-dimensional object detection plays a pivotal role in scene understanding and holds significant importance in various indoor perception applications. Traditional methods based on Hough voting are susceptible to interference from background points or neighboring objects when casting votes for the target’s center from each seed point. Moreover, fixed-size set abstraction modules may result in the loss of structural information for large objects. To address these challenges, this paper proposes a three-dimensional object detection model based on seed point offset attention. The objective of this model is to enhance the model’s resilience to voting noise interference and alleviate feature loss for large-scale objects. Specifically, a seed point offset tensor is first defined, and then the offset tensor self-attention network is employed to learn the weights between votes, thereby establishing a correlation between the voting semantic features and the object structural information. Furthermore, an object surface perception module is introduced, which incorporates detailed features of local object surfaces into global feature representations through vote backtracking and surface mapping. Experimental results indicate that the model achieved excellent performance on the ScanNet-V2 ([email protected], 60.3%) and SUN RGB-D ([email protected], 64.0%) datasets, respectively improving by 2.6% ([email protected]) and 5.4% ([email protected]) compared to VoteNet.

The 3D packing problem has a wide range of applications. However, the complex geometry of irregular objects leads to a sharp increase in the number of placement combinations, making it a challenging problem. In this paper, we propose a packing pipeline based on rigid body dynamics simulation to deal with two types of 3D packing problems. One is the variant bin packing problem, which involves placing more objects into a container of given dimensions to maximize space utilization. The other is the open dimension problem, where the goal is to minimize the container that can accommodate all objects. We first use heuristic placement strategies and a fast collision detection algorithm to efficiently obtain initial packing results. Then, we simulate the shaking of the container according to the dynamic principle. Combined with the vacant space filling operation, shaking the container drives the movement of objects in the container to make the arrangement of objects more compact. For the open dimension packing, the container height is optimized by adjusting the constraints of simulation in the basic pipeline. Experimental results show that our method has advantages over existing methods in both speed and packing density.

Foundation models, such as ULIP-2 (Xue et al., 2023) recently projected forward the field of 3D deep learning. These models are trained with significantly more data and show superior representation learning capacity in many downstream tasks like 3D shape classification and few-shot part segmentation.

A particular characteristic of the recent 3D foundation models is that they are typically multi-modal, and involve image (2D) as well as caption (text) branches. This leads to an intricate interplay that benefits all modalities. At the same time, the nature of the 3D encoders alone, involved in these foundation models is not well-understood. Specifically, there is little analysis on the utility of both pre-trained 3D features provided by these models, or their capacity to adapt to new downstream 3D data. Furthermore, existing studies typically focus on label-oriented downstream tasks, such as shape classification, and ignore other critical applications, such as 3D content-based object retrieval.

In this paper, we fill this gap and show, for the first time, how 3D foundation models can be leveraged for strong 3D-to-3D retrieval performance on seven different datasets, on par with state-of-the-art view-based architectures. We evaluate both the pre-trained foundation models, as well as their fine-tuned versions using downstream data. We compare supervised fine-tuning using classification labels against two self-supervised label-free fine-tuning methods. Importantly, we introduce and describe a methodology for fine-tuning, as we found this to be crucial to make transfer learning from 3D foundation models work in a stable manner.